This is a library provided in the form of a bundler extension that allows the conversion of Dataflow notebooks into their IPykernel equivalents. There is no guarantee made that the Notebook that enters will be as efficient as the notebook in the Dfkernel but it will perform in the same way.
A topological sort if also applied to the Notebook to ensure that it can be ran top down.
It relies on IPython core methods for some of the translation process so some magic and system commands may be translated into their IPython equivalent.
- cd to outer
dfconvertthat containssetup.py. pip install .jupyter bundlerextension enable --sys-prefix --py dfconvert
By enabling the bundler extension you will have the option inside of the File -> Download As method which provides two functions Ipykernel Compatible Notebook.
Optionally the package can also be called by the use of
import dfconvert.make_ipy as ipy
file_name = 'mynotebook.ipynb'
nb = nbformat.read(file_name,nbformat.NO_CONVERT)
new_file_name = 'mynewnotebook.ipynb'
dfpy.export_dfpynb(nb, in_fname=file_name, out_fname=new_file_name, md_above=True,full_transform=False,out_mode=False)
This will create a notebook with out_fname as an ipykernel compatible notebook if out_fname is not set then a file will be created with file_name that includes a _dfpy before the extension to ensure that files do not become overwritten. md_above is a flag for marking markdown cells at the top of the notebook and sending the full_transform flag set to true will also convert all Out[aaaaaa] style references including those that are inside of comments and inside of strings, by default this set to off. out_mode is an additional flag that will make sure that if a cell has any output that is normally created in the dfkernel that this output will be shown in a new cell in the ipykernel, this will be repeated for all results found in a tuple.